ai apis / compared / token cost
What an LLM API really costs
200,000 requests a month to a model cost $0.00 a month, or $0 a year.
list prices, read from the providers' pages · checked 17 november 2026 · sources
model provider
a year $0
you can keep
$0.00 a month
That is what the Batch API would save on this model: a month instead, for work that can wait up to a day.
Why the bill is bigger than expected
Compare on your own token mix. Prices per million tokens differ by model and by direction. A model that is cheap on input can be dear on output: the bill depends on how much you send and how much it writes back.
Caching and batch change the ranking. Cached input costs a tenth of the normal price at Anthropic, OpenAI and Google, and their batch APIs cost half. A workload with a long shared prompt or no urgency can cost far less than the list price suggests.
Count the tokens, not the words. A token is about three quarters of an English word. Code, other languages and structured data use more tokens per word.
Sources
- [1]Anthropic API prices per million tokens, standard, cached and batch, as published. docs.claude.com/en/docs/about-claude/pricing, read 2026-11-17.
- [2]OpenAI API prices per million tokens, standard, cached and batch, as published. platform.openai.com/docs/pricing, read 2026-11-17.
- [3]Google API prices per million tokens, standard, cached and batch, as published. ai.google.dev/gemini-api/docs/pricing, read 2026-11-17.
Prices are list prices in USD per million tokens, before tax, credits or committed-use discounts. They are kept by hand from each provider's pricing page and confirmed against it every night; a price that no longer matches blocks the site from updating. Cache writes, tools such as web search, images, audio and fine-tuning are not counted. The calculator runs in your browser; nothing you type is sent anywhere.